This document discusses how human and machine intelligence can work together in data science. It addresses key areas like data, models, and decisions. For data, it emphasizes representativeness and avoiding biases. For models, it advocates bringing the best of human and machine through techniques like human-in-the-loop machine learning. For decisions, it notes the importance of considering costs of mistakes, knowing limits of models, and prototyping systems before full deployment. The overall message is that close collaboration between human and AI experts can help ensure systems are effective, fair and accountable.
1. PEOPLE + AI
“Ta” Virot Chiraphadhanakul
virot.ch@skooldio.com
Human intelligence meets machine intelligence
2. Virot Chiraphadhanakul
วิโรจน์ จิรพัฒนกุล (ต้า)
B.Eng. (Computer), Chulalongkorn
Ph.D. in Operations Research, MIT
ex-Data Scientist, Facebook
Google Developers Expert in Machine Learning
Lecturer @ Chula Business School
Managing Director @ Skooldio
4. “I believe in a world that will
have an abundance of artificial
intelligence, but what will be
scarce is real intelligence and
human qualities, like empathy”
— Satya Nadella, Microsoft CEO
20. “The problems may be due to widespread
media coverage of this year’s severe US flu
season […] The press reports may have
triggered many flu-related searches by
people who were not ill.”
When
Google got
flu wrong
Source: http://www.nature.com/news/when-google-got-flu-wrong-1.12413
25. “the decisions you make today
affect the training data you collect
tomorrow. Once your machine
learning system embeds biases
into its model, it can continue
generating new training data that
reinforces those biases.”
Source: https://medium.com/@dtunkelang/10-things-everyone-should-know-about-machine-learning-15279c27ce96
45. • Consider the following scenarios
• A customer is wrongly alerted
(false positive)
• A customer is wrongly charged for
a fraudulent transaction (false
negative)
Unequal
Costs and
Benefits